components/dom_distiller/core/distillable_page_detector.cc

   1 // Copyright 2015 The Chromium Authors. All rights reserved.
   2 // Use of this source code is governed by a BSD-style license that can be
   3 // found in the LICENSE file.
   4
   5 #include "components/dom_distiller/core/distillable_page_detector.h"
   6
   7 #include "base/logging.h"
   8
   9 namespace dom_distiller {
  10
  11 DistillablePageDetector::DistillablePageDetector(
  12     scoped_ptr<AdaBoostProto> proto)
  13     : proto_(proto.Pass()), threshold_(0.0) {
  14   CHECK(proto_->num_stumps() == proto_->stump_size());
  15   for (int i = 0; i < proto_->num_stumps(); ++i) {
  16     const StumpProto& stump = proto_->stump(i);
  17     CHECK(stump.feature_number() >= 0);
  18     CHECK(stump.feature_number() < proto_->num_features());
  19     threshold_ += stump.weight() / 2.0;
  20   }
  21 }
  22
  23 DistillablePageDetector::~DistillablePageDetector() {
  24 }
  25
  26 bool DistillablePageDetector::Classify(
  27     const std::vector<double>& features) const {
  28   return Score(features) > threshold_;
  29 }
  30
  31 double DistillablePageDetector::Score(
  32     const std::vector<double>& features) const {
  33   CHECK(features.size() == size_t(proto_->num_features()));
  34   double score = 0.0;
  35   for (int i = 0; i < proto_->num_stumps(); ++i) {
  36     const StumpProto& stump = proto_->stump(i);
  37     if (features[stump.feature_number()] > stump.split()) {
  38       score += stump.weight();
  39     }
  40   }
  41   return score;
  42 }
  43
  44 double DistillablePageDetector::GetThreshold() const {
  45   return threshold_;
  46 }
  47
  48 }  // namespace dom_distiller